Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network

David R. Chambers; Bradley B. Brimhall; Donald R. Poole; Edward A. Medina

doi:10.1117/12.2611636

Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network

David R. Chambers, Bradley B. Brimhall, Donald R. Poole, Edward A. Medina

Department of Pathology

Producción científica: Conference contribution

2 Citas (Scopus)

Resumen

In our work, we present an approach to regressing breast cancer cellularity in patches extracted from Whole Slide Imagery (WSI) on Hematoxylin and Eosin (H&E) stains using a fully-convolutional neural network which is trained with two heads: one which computes a global average pool for weakly-labeled data (data with a cellularity score of 0- 1.0) and another which enforces pixel-wise activations for strongly-labeled (segmentation) data. Our method was the top-performing algorithm of all submissions to the BreastPathQ challenge, achieving a prediction probability of 0.941.

Idioma original	English (US)
Título de la publicación alojada	Medical Imaging 2022
Subtítulo de la publicación alojada	Digital and Computational Pathology
Editores	John E. Tomaszewski, Aaron D. Ward, Richard M. Levenson
Editorial	SPIE
ISBN (versión digital)	9781510649538
DOI	https://doi.org/10.1117/12.2611636
Estado	Published - 2022
Evento	Medical Imaging 2022: Digital and Computational Pathology - Virtual, Online Duración: mar 21 2022 → mar 27 2022

Serie de la publicación

Nombre	Progress in Biomedical Optics and Imaging - Proceedings of SPIE
Volumen	12039
ISSN (versión impresa)	1605-7422

Conference

Conference	Medical Imaging 2022: Digital and Computational Pathology
Ciudad	Virtual, Online
Período	3/21/22 → 3/27/22

ASJC Scopus subject areas

Electronic, Optical and Magnetic Materials
Atomic and Molecular Physics, and Optics
Biomaterials
Radiology Nuclear Medicine and imaging

Acceder al documento

10.1117/12.2611636

Otros archivos y enlaces

Citar esto

Chambers, D. R., Brimhall, B. B., Poole, D. R., & Medina, E. A. (2022). Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network. En J. E. Tomaszewski, A. D. Ward, & R. M. Levenson (Eds.), Medical Imaging 2022: Digital and Computational Pathology Artículo 120390A (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 12039). SPIE. https://doi.org/10.1117/12.2611636

Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network. / Chambers, David R.; Brimhall, Bradley B.; Poole, Donald R. et al.
Medical Imaging 2022: Digital and Computational Pathology. ed. / John E. Tomaszewski; Aaron D. Ward; Richard M. Levenson. SPIE, 2022. 120390A (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 12039).

Producción científica: Conference contribution

Chambers, DR, Brimhall, BB, Poole, DR & Medina, EA 2022, Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network. En JE Tomaszewski, AD Ward & RM Levenson (eds.), Medical Imaging 2022: Digital and Computational Pathology., 120390A, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 12039, SPIE, Medical Imaging 2022: Digital and Computational Pathology, Virtual, Online, 3/21/22. https://doi.org/10.1117/12.2611636

Chambers DR, Brimhall BB, Poole DR, Medina EA. Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network. En Tomaszewski JE, Ward AD, Levenson RM, editores, Medical Imaging 2022: Digital and Computational Pathology. SPIE. 2022. 120390A. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE). doi: 10.1117/12.2611636

Chambers, David R. ; Brimhall, Bradley B. ; Poole, Donald R. et al. / Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network. Medical Imaging 2022: Digital and Computational Pathology. editor / John E. Tomaszewski ; Aaron D. Ward ; Richard M. Levenson. SPIE, 2022. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE).

@inproceedings{2472b9cb084f47e498b779d68e5394a9,

title = "Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network",

abstract = "In our work, we present an approach to regressing breast cancer cellularity in patches extracted from Whole Slide Imagery (WSI) on Hematoxylin and Eosin (H\&E) stains using a fully-convolutional neural network which is trained with two heads: one which computes a global average pool for weakly-labeled data (data with a cellularity score of 0- 1.0) and another which enforces pixel-wise activations for strongly-labeled (segmentation) data. Our method was the top-performing algorithm of all submissions to the BreastPathQ challenge, achieving a prediction probability of 0.941.",

keywords = "Digital pathology, cancer cellularity, convolutional neural networks, whole slide imagery",

author = "Chambers, \{David R.\} and Brimhall, \{Bradley B.\} and Poole, \{Donald R.\} and Medina, \{Edward A.\}",

note = "Publisher Copyright: {\textcopyright} COPYRIGHT SPIE.; Medical Imaging 2022: Digital and Computational Pathology ; Conference date: 21-03-2022 Through 27-03-2022",

year = "2022",

doi = "10.1117/12.2611636",

language = "English (US)",

series = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",

publisher = "SPIE",

editor = "Tomaszewski, \{John E.\} and Ward, \{Aaron D.\} and Levenson, \{Richard M.\}",

booktitle = "Medical Imaging 2022",

}

TY - GEN

T1 - Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network

AU - Chambers, David R.

AU - Brimhall, Bradley B.

AU - Poole, Donald R.

AU - Medina, Edward A.

N1 - Publisher Copyright: © COPYRIGHT SPIE.

PY - 2022

Y1 - 2022

N2 - In our work, we present an approach to regressing breast cancer cellularity in patches extracted from Whole Slide Imagery (WSI) on Hematoxylin and Eosin (H&E) stains using a fully-convolutional neural network which is trained with two heads: one which computes a global average pool for weakly-labeled data (data with a cellularity score of 0- 1.0) and another which enforces pixel-wise activations for strongly-labeled (segmentation) data. Our method was the top-performing algorithm of all submissions to the BreastPathQ challenge, achieving a prediction probability of 0.941.

AB - In our work, we present an approach to regressing breast cancer cellularity in patches extracted from Whole Slide Imagery (WSI) on Hematoxylin and Eosin (H&E) stains using a fully-convolutional neural network which is trained with two heads: one which computes a global average pool for weakly-labeled data (data with a cellularity score of 0- 1.0) and another which enforces pixel-wise activations for strongly-labeled (segmentation) data. Our method was the top-performing algorithm of all submissions to the BreastPathQ challenge, achieving a prediction probability of 0.941.

KW - Digital pathology

KW - cancer cellularity

KW - convolutional neural networks

KW - whole slide imagery

UR - https://www.scopus.com/pages/publications/85132835974

UR - https://www.scopus.com/inward/citedby.url?scp=85132835974&partnerID=8YFLogxK

U2 - 10.1117/12.2611636

DO - 10.1117/12.2611636

M3 - Conference contribution

AN - SCOPUS:85132835974

T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

BT - Medical Imaging 2022

A2 - Tomaszewski, John E.

A2 - Ward, Aaron D.

A2 - Levenson, Richard M.

PB - SPIE

T2 - Medical Imaging 2022: Digital and Computational Pathology

Y2 - 21 March 2022 through 27 March 2022

ER -

Cancer cell segmentation for cellularity prediction via a weakly labeled/strongly labeled hybrid convolutional neural network

Resumen

Serie de la publicación

Conference

ASJC Scopus subject areas

Acceder al documento

Otros archivos y enlaces

Huella

Citar esto